Screening, Devices, and Digital Fraud

A screening device does not have to prove what a stone is; its task can be only to safely separate items that pass a defined filter from those that require additional examination. Digital fraud further demonstrates why the device result, the document, and the physical stone must remain separate layers.

Essential

A handheld property tester and a Diamond Verification Instrument (DVI) are not the same category. Screening systems often use results such as PASS, REFER, or DETECT, but the meaning is valid only within the declared scope of the device.

REFER is not a synonym for synthetic (laboratory-grown). It means that the device did not produce a result that permits the workflow to end and that a next step is required.

Loose, mounted, and melee material can require separately validated procedures. Throughput is processing speed, not accuracy.

ASSURE as an Example of Validation

ASSURE 2.0 tests devices on a standardized challenge sample. Such a sample is used to compare performance, not to estimate market prevalence of individual stone types.

False Positive Rate, Referral Rate, and Accuracy are not the same metric. Even a result of 0% false positives in one test is not a universal promise that the device will never make an error.

For every performance claim, record the model, version, sample, test protocol, and date.

Digital Fraud

Three questions must remain separate:

1. Is the physical stone what it is claimed to be?
2. Is the document authentic?
3. Does that document correspond to that specific stone?

A counterfeit laser inscription does not by itself determine natural/LGD/treatment status. A near-match stone can deliberately imitate the weight, dimensions, and grade of a legitimate report.

A report number that exists in an official database confirms the record, but not the physical match. A PDF, screenshot, or QR destination is a weaker evidentiary layer than the official record.

Weight and dimensions are not an individual fingerprint. A laser inscription is useful, but it is on the surface and can be changed. A spectroscopic or imaging mismatch can reveal a substitution that a surface comparison does not detect.

How to Proceed

For an important transaction:

• check the official record;
• inspect the inscription if one exists;
• compare weight, measurements, plot/inclusions, and other characteristics;
• use screening only within the scope of the device;
• in the event of inconsistency or high risk, obtain an independent laboratory re-examination.

Generative AI increases the need for source verification. A document that “looks perfect” does not become authentic because it is visually convincing.

Practical Framework: Screening Is Workflow Management

A screening system exists to quickly separate samples that can be passed forward within the validated scope from those that need additional analysis. Labels such as PASS, REFER, DETECT, or OUT-OF-RANGE are therefore meaningful only with documentation for the specific device and the population on which it was validated.

Refer is not a verdict. It can mean that the sample shows a signal the device cannot resolve safely, that it falls outside the configuration for which the method was validated, or simply that another analytical step is required. In a professional system, referral rate and false-positive rate are not the same metric, and performance on a challenge set such as an ASSURE test should not be converted into a claim about the actual prevalence of synthetic material in the market. (In the United States, the FTC Jewelry Guides, 16 CFR Part 23, also permit the qualifiers laboratory-grown and laboratory-created; jurisdiction: US.)

Digital fraud adds another layer. An authentic report number can exist while the listing or physical stone still does not correspond to that record. Near-match dimensions, copied photographs, a copied PDF, or a counterfeit inscription require verification of the record and physical matching as two separate procedures.

For an operational workflow, document the device/version, scope, serial number or calibration status where relevant, result, and next step. For a melee parcel, separate population screening from identification of an individual stone. The goal is not for the device to “get everything right,” but to manage risk and escalation reliably.

When to Escalate

Escalate every REFER or OUT-OF-RANGE result to a method capable of resolving the remaining gemological question. When digital fraud is suspected, independently open the laboratory’s official channel, verify the original record, and then establish the physical match to the stone; a PDF, screenshot, or QR code alone proves neither authenticity nor physical identity.

Quick Check Before Reaching a Conclusion

Before accepting a technical, purchasing, or documentary conclusion, run this short check:

• Do I know the device, version, and validated scope of the screening system?
• Am I distinguishing referral rate, false positive, and actual prevalence?
• Am I interpreting a challenge-set result without turning it into market statistics?
• Am I checking the official record and physical stone as two separate layers?
• Am I documenting the result and a predefined next step for every REFER/OUT-OF-RANGE?

Common Mistakes

“REFER means synthetic.”
No.

“0% false positives means an infallible device.”
Not outside the specific protocol that was tested.

“The report exists online, so this must be the stone.”
No. The record and physical match are separate.

“A QR code proves authenticity.”
Not by itself.

Remember

Screening is triage, not magic. Digital verification is layered: device + official record + physical match + exception handling.

Go Deeper in The Book

• Chapter 73 — Screening Devices and the ASSURE Program
• Chapter 74 — Counterfeiting and Digital Fraud